Abstract
Background
In resource limited settings clinical parameters, including body weight changes, are used to monitor clinical response. Therefore we studied body weight changes in patients on antiretroviral treatment (ART) in different regions of the world.
Methods
Data were extracted from the “International Epidemiologic Databases to Evaluate AIDS”, a network of ART programmes that prospectively collects routine clinical data. Adults on ART from the Southern-, East-, West- and Central African and the Asia-Pacific regions were selected from the database if baseline data on body weight, gender, ART regimen and CD4 count were available. Body weight change over the first two years and the probability of body weight loss in the second year were modelled using linear mixed models and logistic regression respectively.
Results
Data from 205,571 patients were analysed. Mean adjusted body weight change in the first 12 months was higher in patients started on tenofovir and/or efavirenz; in patients from Central, West and East Africa, in men, and in patients with a poorer clinical status. In the second year of ART it was greater in patients initiated on tenofovir and/or nevirapine, and for patients not on stavudine, in women, in Southern Africa and in patients with a better clinical status at initiation. Stavudine in the initial regimen was associated with a lower mean adjusted body weight change and with weight loss in the second treatment year.
Conclusion
Different ART regimens have different effects on body weight change. Body weight loss after one year of treatment in patients on stavudine might be associated with lipoatrophy.
Keywords: HIV, bodyweight, determinants, ART, low-resource, settings
Introduction
Monitoring HIV-positive patients on antiretroviral treatment (ART) remains a challenge in countries with limited resources. In many of them routine viral load and even CD4 lymphocyte count monitoring is not possible or available.1 Therefore in these settings clinical parameters, including body weight changes, are used to monitor clinical response.2
Significant body weight loss before the start of ART is usually associated with advancing HIV disease and the presence of opportunistic infections (in low resource countries mainly tuberculosis).3 Short term weight evolution is an indicator of treatment success. Indeed, weight gain at 3 months was found to be strongly associated with survival4 and weight loss as early as 1 to 6 months after ART initiation has been associated with a high risk of adverse outcomes.5, 6, 7 Prior studies mainly focused on early periods during ART (generally less than 12 months).
The objective of our study was to describe body weight changes and associations with those changes across regions of the world among patients receiving up to two years of ART.
Methods
Data were extracted from the “International Epidemiologic Databases to Evaluate AIDS”, a large network of ART programmes that prospectively collects routine clinical data (IeDEA) (http://www.iedea.org/). Its aim is to collect baseline and follow-up characteristics on HIV-positive patients initiating ART worldwide.8 These characteristics include demographic (age, gender, geographical region), clinical (weight, height, pregnancies, morbidity, medication) and biological (CD4 count, viral load) data. Adult patients (>18 years) on ART from the Southern-, East-, West- and Central African regions and the Asia-Pacific region were selected from the database if baseline data (within a 60 day window before start of ART) on body weight, gender, ART regimen and CD4 count were available. Data for the study were collected from 140 sites, including 18 from Asia-Pacific, 10 from Central-Africa, 10 from East-Africa, 87 from Southern Africa and 15 from West Africa. HIV-positive patients were included in this analysis if at ART initiation they were ART naïve and starting first line ART containing at least 3 different antiretroviral drugs. Pregnant women and patients on investigational study drugs or on implausible regimens (regimens containing the following combinations: zidovudine (AZT) and stavudine (D4T), AZT and tenofovir (TDF), D4T and TDF, efavirenz (EFV) and nevirapine (NVP)) were excluded. All patients that had at least one baseline weight measure were included. Body weight change at time t is defined as body weight at time t minus body weight at ART initiation.
Statistical analysis
The body mass index (BMI) was computed when body weight and height were both available; its 24-month evolution was graphically represented by region. As a significant proportion of patients had missing height, the main analytical models were based on body weight instead of BMI to avoid a selection bias. Individuals with complete and incomplete BMI data were compared.
Model 1
Body weight change was modeled over the first two years on ART using linear mixed models (LMM) with no intercept and two slopes; the first slope over the first year of ART and the second slope over the second. To account for intra-individual correlation, we added random effects on the two slopes with an unstructured variance-covariance matrix. The LMMs were adjusted for geographical region, gender, age, initial body weight, initial clinical stage, first ART regimen, initial hemoglobin, calendar year of ART initiation and initial CD4 count. The first body weight change slope was also adjusted for CD4 count changes between month 0 and month 12 and the second slope was adjusted for CD4 count change between 12 and 24 months. Moreover, we let the association between D4T and body weight change in the second year of ART to interact with baseline body weight, region, gender, age at ART initiation and baseline CD4 count.
Model 2
This model studied risk factors for any weight loss larger than 5% in the second year of ART by fitting a multiple logistic model with weight loss larger than 5% during the second year as a binary outcome variable. Not all patients had a weight measurement exactly at 1 year and 2 years after start of ART treatment. Hence we estimated the weight after the first year as the mean weight between 6 and 18 months and the weight after the second year as the mean weight between 18 and 30 months.
To account for missing data, missing CD4 counts were imputed using CD4 counts estimated with a predictive LMM, adjusted for geographical region, gender, age, initial body weight, clinical stage, ART and initial hemoglobin.
Results
Baseline characteristics
Data from 212,795 patients were received from the IeDEA regions. Reasons for exclusion of 7,224 (3.4%) patients were: age under 18 years (n=581) and implausible regimen (n=6643) leaving 205,571 patients for analysis (139,174 (67.7%) from Southern Africa, 42,856 (20.8%) from East Africa, 17,202 (8.4%) from West Africa, 4,700 (2.3%) from Central Africa and 1,639 (0.8%) from Asia Pacific) (See supplemental table 1 for number of patients per country). Patient characteristics at ART initiation by geographical region are described in table 1. Of the 205,571 patients included in the analysis of the first year of ART in 58,835 (28.6%) of them only a weight measurement at baseline was available. In the remaining patients the median number of weight measurements was 3 (IQR 2–4). In the second year of ART 104,744 patients had at least 1 weight measurement. The median number of weight measurements was 3 (IQR 2–5). Patients initiated ART between the years 2001 and 2010. The median (Inter-Quartile Range [IQR]) body weight at ART initiation was 55 kg [48–62], 58 kg [52–65] at 6 months, 60 kg [53–67] at 12 months and 60 kg [54–68] at 24 months on ART (Table 1).
Table 1.
Patient characteristics at antiretroviral therapy (ART) initiation by geographical region (n=205,571). International Epidemiologic Databases to Evaluate AIDS (IeDEA)
| Asia Pacific | Central Africa | East Africa | Southern Africa | West Africa | Total | ||
|---|---|---|---|---|---|---|---|
| N (%) | 1,639 (0.8) | 4,700 (2.3) | 42,856 (20.8) | 139,174 (67.7) | 17,202 (8.4) | 205,571 (100) | |
| Year of enrolment (range) | 1994–2010 | 1998–2010 | 2001–2009 | 1997–2010 | 1998–2009 | 1994–2010 | |
| Female (%) | 588 (35.9) | 3,297 (70.1) | 27,106 (63.2) | 84,078 (60.4) | 11,075 (64.4) | 126,144 (61.4) | |
| Age (years) median (IQR) | 34 (29–40) | 38 (31–44) | 37 (31–43) | 35 (30–42) | 36 (30–43) | 35 (30–42) | |
| Clinical stage (%) | |||||||
| WHO¥ I/II or CDC≥ A | 623 (38.0) | 1,413 (30.1) | 13,263 (30.9) | 45,194 (32.5) | 6,229 (36.2) | 66,722 (32.5) | |
| WHO III or CDC B | 455 (27.8) | 2,615 (55.6) | 18,383 (42.9) | 68,408 (49.2) | 7,165 (41.7) | 97,026 (47.2) | |
| WHO IV or AIDS | 561 (34.2) | 652 (13.9) | 6,442 (15.0) | 12,144 (8.7) | 1,795 (10.4) | 21,594 (10.5) | |
| Missing | - | 20 (0.4) | 4,768 (11.1) | 13,428 (9.6) | 2,013 (11.7) | 20,229 (9.8) | |
| Year of ART initiation (%) | |||||||
| <2005 | 655 (40.0) | 463 (9.9) | 3,070 (7.2) | 7,063 (5.1) | 2,741 (15.9) | 13,992 (6.8) | |
| 2005–2006 | 567 (34.6) | 870 (18.5) | 19,321 (45.1) | 35,178 (25.3) | 8,026 (46.7) | 63,962 (31.1) | |
| 2007–2008 | 302 (18.4) | 2,094 (44.6) | 20,085 (46.9) | 53,071 (38.1) | 6,219 (36.2) | 81,771 (39.8) | |
| 2009–2010 | 115 (7.0) | 1,273 (27.1) | 380 (7.2) | 43,862 (31.5) | 216 (1.3) | 45,846 (22.3) | |
| First ART regimen, NRTI^ (%) | |||||||
| AZT-based | 598 (36.5) | 1,857 (39.5) | 9,688 (22.6) | 28,177 (20.2) | 5,751 (33.4) | 46,071 (22.4) | |
| D4T-based | 953 (58.1) | 2,752 (58.6) | 32,928 (76.8) | 64,265 (46.2) | 9,320 (54.2) | 110,218 (53.6) | |
| Tenofovir-based | 56 (3.4) | 66 (1.4) | 225 (0.5) | 43,515 (31.3) | 1,872 (10.9) | 45,734 (22.2) | |
| Other NRTI’s* | 32 (2.0) | 25 (0.5) | 15 (0.0) | 3,217 (2.3) | 259 (1.5) | 3,548 (1.7) | |
| First ART regimen, NNRTI# or PI (%) | |||||||
| Efavirenz-based | 487 (29.7) | 715 (15.2) | 8,035 (18.7) | 56,287 (40.4) | 5,514 (32.1) | 71,038 (34.6) | |
| Nevirapine-based | 998 (60.9) | 3,933 (83.7) | 34,724 (81.0) | 82,314 (59.1) | 10,119 (58.8) | 132,088 (64.3) | |
| Protease inhibitor | 154 (9.4) | 52 (1.1) | 97 (0.2) | 573 (0.4) | 1,569 (9.1) | 2,445 (1.2) | |
| CD4 count in cells/µl median (IQR) | 95 (33–188) | 135 (65–204) | 102 (40–175) | 134 (67–202) | 139 (60–223) | 128 (61–199) | |
| CD4 count missing (%) | 341 (20.8) | 1,316 (28.0) | 16,212 (37.8) | 33,497 (24.1) | 4,312 (25.1) | 55,678 (27.1) | |
| Hemoglobin in g/dl median (IQRε) | 12.1 (10.9–13.5) | 10.2 (9.0–12.0) | 11.2 (9.6–12.8) | 10.9 (9.4–12.4) | 10.2 (9.0–11.5) | 10.9 (9.4–12.4) | |
| Hemoglobin missing (%) | 1,314 (80.2) | 3,966 (84.4) | 23,957 (55.9) | 39,770 (28.6) | 6,203 (36.1) | 75,310 (36.6) | |
| Body weight in kg, median (IQR) | |||||||
| At ART initiation (IQR) | 54 (47–61) | 56 (49–64) | 54 (48–61) | 55 (48–62) | 56 (49–64) | 55 (48–62) | |
| At 6 months (IQR) | 56 (50–65) | 59 (53–68) | 59 (53–65) | 58 (52–65) | 60 (53–68) | 58 (52–65) | |
| At 12 months (IQR) | 57 (51–65) | 61 (54–70) | 60 (54–67) | 59 (53–67) | 62 (55–70) | 60 (53–67) | |
| At 24 months (IQR) | 57 (50–64) | 62 (54–71) | 60 (54–67) | 60 (54–68) | 62 (55–71) | 60 (54–68) | |
| Body mass index in kg/m2, median (IQR) | |||||||
| At ART initiation (IQR) | 20.1 (18.1–22.5) | 20.6 (18.2–23.5) | 19.7 (17.6–21.9) | 20.0 (17.9–22.5) | 20.6 (18.2–23.4) | 20.0 (17.9–22.5) | |
| At 6 months (IQR) | 21.3 (19.5–23.7) | 22.0 (19.8–24.7) | 21.2 (19.4–23.5) | 21.4 (19.5–23.9) | 22.1 (20.0–24.9) | 21.4 (19.5–23.9) | |
| At 12 months (IQR) | 21.7 (19.8–23.8) | 22.6 (20.3–25.7) | 21.7 (19.7–24.1) | 21.9 (19.8–24.6) | 22.8 (20.4–25.6) | 21.9 (19.9–24.6) | |
| At 24 months (IQR) | 21.6 (19.8–23.7) | 23.0 (20.3–26.2) | 21.7 (19.7–24.1) | 22.1 (19.9–24.9) | 23.0 (20.5–26.0) | 22.1 (19.9–24.8) | |
WHO: World Health Organisation
CDC: Centers for Disease Control
NRTI: Nucleoside/Nucleotide reverse transcriptase inhibitor
Other NRTI’s: didanosine, lamivudine, abacavir and emtricitabine
NNRTI: Non–nucleoside reverse transciptase inhibitor
PI : protease inhibitor
IQR : Inter–quartile range
In all African regions, in contrast to the Asia-Pacific region, more women than men were enrolled in the cohorts (table 1). The median CD4 lymphocyte count at ART initiation was 128 cells/µL (IQR 61–199) and the median CD4 lymphocyte count was below 150 cells/µL in all cohorts. Most patients (57.7%) were in WHO stage III or IV. Over half of regimens contained D4T (n=110,218, 53.6%) and roughly two-thirds contained NVP (n=132,088, 64.3%). Only in Southern Africa a relatively large proportion of patients (n=43,515, 31.3%) were started on a TDF-containing regimen (table 1).
BMI evolution in different regions
BMI was computed for 165,804 individuals at baseline (80.7% of the sample), 89,762 at M6 (43.7%), 70,285 at M12 (34.2%), 56,350 at M18 (27.4%) and 44,598 at M24 (21.7%). Individuals with complete and incomplete BMI data were comparable at baseline (Supplemental table 2).
BMI evolution within the first 2 years of ART per region is shown in Figure 1. For all regions, median BMI increased in the first 12 months after ART initiation, after which it levelled off. Obesity (BMI>30 kg/m2) was present in 4402 patients (2.6%) at baseline and in 2644 (5.9%) after 2 years on ART. Of those patients with a BMI>30 kg/m2 after 2 years on ART, the BMI at baseline was ≤18 in 40 (1.5%), between 18 and 25 in 705 (26.7%), between 25 and 30 in 1016 (38.4%) and >30 kg/m2 in 883 (33.4%).
Figure 1.
Evolution of body mass index within the first 2 years of ART per region in the study population (n=165,804). International Epidemiologic Databases to Evaluate AIDS (IeDEA)
Body weight change after ART initiation
Adjusted body weight change slopes (kg/year) for all patients from 0 to 12 months and from 12 to 24 months after ART initiation (Model 1) are shown in Table 2, taking as a reference in both analyses men aged <30 years who started ART in Southern Africa with TDF and EFV in 2009 or later, with an initial clinical stage WHO IV (or AIDS), with a baseline CD4 count <50 cells/µL, a CD4 cell count change between 0 and 149 at month 12 and between 0 and 49 at month 24, hemoglobinemia <7.5 g/dL and with an initial body weight of 55 kg.
Table 2.
Body weight change slopes within the first 24 months after ART initiation modelled with the adjusted Linear Mixed Model: main effects (n=205,571; observations=1,785,439)
| From Month 0 to 12 | From Month 12 to 24 | |||
|---|---|---|---|---|
| Body weight change in kg/year |
95% confidence interval |
Body weight change in kg/year |
95% confidence interval |
|
| Body weight change in reference group‡ | 9.12 | 8.81; 9.42 | −0.08 | −0.50;0.34 |
| Gender (versus ref‡) | ||||
| Men | Ref‡ | Ref‡ | ||
| Women | −0.29 | −0.38; −0.20 | 1.22 | 1.09;1.35 |
| NRTI^ in first ART regimen (versus ref‡) | ||||
| Tenofovir (TDF) | Ref‡ | Ref‡ | ||
| Stavudine (D4T) | −0.16 | −0.29; −0.02 | 0.01 | −0.30;0.32 |
| Zidovudine (AZT) | −0.29 | −0.43; −0.15 | −0.45 | −0.61;−0.28 |
| Other NRTIs* | −0.35 | −0.70; −0.004 | −0.59 | −1.09;−0.09 |
| NNRTI# or PIß in first ART regimen (versus ref‡) | ||||
| Efavirenz (EFV) | Ref‡ | Ref‡ | ||
| Nevirapine (NVP) | −0.69 | −0.78; −0.59 | 0.34 | 0.23;0.44 |
| Protease inhibitors | −1.41 | −1.79; −1.03 | −0.25 | −0.62;0.12 |
| Geographical region (versus ref‡) | ||||
| Southern Africa | Ref‡ | Ref‡ | ||
| Asia-Pacific | −0.41 | −0.81; −0.01 | −0.68 | −1.23;−0.12 |
| Central Africa | 1.85 | 1.54; 2.16 | −0.63 | −1.19;−0.07 |
| East Africa | 0.46 | 0.35; 0.56 | −0.28 | −0.53;−0.03 |
| West Africa | 1.56 | 1.40; 1.71 | −0.08 | −0.30;0.15 |
| Year of ART initiation (versus ref‡) | ||||
| 2009–2010 | Ref‡ | Ref‡ | ||
| 2007–2008 | 0.56 | 0.43; 0.69 | −1.46 | −1.72; −1.21 |
| 2005–2006 | 1.45 | 1.31; 1.60 | −1.43 | −1.70; −1.16 |
| ART <2005 | 1.56 | 1.36; 1.75 | −0.93 | −1.22; −0.64 |
| Age (years) at ART initiation(versus ref‡) | ||||
| 18–29 | Ref‡ | Ref‡ | ||
| 30–34 | 0.41 | 0.30; 0.53 | 0.15 | −0.03;0.33 |
| 35–39 | 0.64 | 0.52; 0.76 | 0.28 | 0.09;0.47 |
| >39 | 0.68 | 0.57; 0.79 | 0.27 | 0.09;0.45 |
| Baseline body weight (per 5 kg)(versus ref‡) | ||||
| 55 kg | Ref‡ | |||
| For each 5kg higher | −0.90 | −0.92; −0.88 | 0.29 | 0.26;0.32 |
| Initial clinical stage (versus ref‡) | ||||
| AIDS or WHO IV | Ref‡ | Ref‡ | ||
| CDC B or WHO III | −0.30 | −0.43; −0.16 | 0.003 | −0.14;0.14 |
| CDC A or WHO I/II | −2.26 | −2.41; −2.12 | 0.40 | 0.26;0.55 |
| Clinical stage missing | −0.49 | −0.68; −0.30 | 0.08 | −0.12;0.28 |
| Baseline CD4 count in cells/µl (versus ref‡) | ||||
| 0–49 | Ref‡ | Ref‡ | ||
| 50–199 | −2.04 | −2.15; −1.92 | 0.08 | −0.11;0.28 |
| >199 | −3.56 | −3.70; −3.42 | 0.24 | 0.02;0.45 |
| CD4 count change between M0 and M12 in cells/µl (versus ref‡) | ||||
| <0 | −0.92 | −1.09; −0.75 | - | |
| 0–149 | Ref‡ | - | ||
| 150–249 | 0.56 | 0.47; 0.65 | - | |
| >249 | 1.43 | 1.33; 1.54 | ||
| CD4 count change between M12 and M24 in cells/µl (versus ref‡) | ||||
| <0 | - | −0.16 | −0.27; −0.04 | |
| 0–49 | - | Ref‡ | ||
| 50–99 | - | −0.06 | −0.17; 0.05 | |
| >99 | - | 0.44 | 0.33; 0.55 | |
| Baseline hemoglobin in g/dl (versus ref‡) | ||||
| <7.5 | Ref‡ | Ref‡ | ||
| 7.5–10 | −0.84 | −1.08; −0.60 | 0.28 | 0.04; 0.52 |
| ≥10 | −2.17 | −2.41; −1.94 | 0.33 | 0.10; 0.56 |
| Missing | −2.36 | −2.60; −2.13 | 0.32 | 0.08; 0.56 |
Reference group: men aged <30 years who started ART in Southern Africa with tenofovir and efavirenz in 2009 or later, with an initial clinical stage WHO IV (or AIDS), with a baseline CD4 count <50 cells/µL, a CD4 count change =0–149 at M12 and =0–49 at M24, hemoglobinemia <7.5 g/dL and with an initial body weight =55 kg.
NRTI: Nucleoside/Nucleotide reverse transcriptase inhibitor
Other NRTI’s: didanosine, lamivudine, abacavir and emtricitabine
NNRTI: Non-nucleoside reverse transciptase inhibitor
PI : protease inhibitor
Gender
In the first year after ART initiation, the mean adjusted body weight change was greater in men than in women (−0.29, 95% CI −0.38; −0.20) but in the second year it was greater in women (1.22, 95% CI1.09;1.35).
ART regimen
Patients initiating TDF-containing regimens had a higher mean adjusted body weight change in the first two years compared to patients who started on a regimen containing AZT or ‘other NRTI’s. Patients started on a D4T-containing regimen had a significant lower mean adjusted body weight change in the first year on ART (−0.16, 95% CI −0.29; −0.02). In the second year on ART there was no significant difference in mean adjusted body weight change between patients on D4T and TDF. However the effect on mean adjusted body weight change in the second year of D4T in initial ART regimen (table 3) was more negative for patients from Eastern Africa (−0.77 kg/year, 95% CI −1.04; −0.50), women (−1.13 kg/year, 95% CI −1.29; −0.96), elder patients (age above 39 years)( −1.16 kg/year, 95% CI −1.38; −0.94), patients with a higher baseline body weight (−0.34 kg/year, 95% CI −0.38; −0.30, for every 5 kg more) and a CD4 cell count below 50 cells/µL at baseline (compared to a baseline CD4 count above 199 cells/µL) (table 3).
Table 3.
Body weight change slopes within the first 24 months after ART initiation modelled with the adjusted Linear Mixed Model: interactions terms with D4T in second year (n=205,571; observations=1,785,439)
| Body weight change in kg/year | 95% confidence interval |
|
|---|---|---|
| Gender (versus ref‡) | ||
| Men | Ref‡ | |
| Women | −1.13 | −1.29;−0.96 |
| Geographical region (versus ref‡) | ||
| Southern Africa | Ref‡ | |
| Asia-Pacific | −0.59 | −1.24;0.07 |
| Central Africa | 1.86 | 1.21;2.51 |
| East Africa | −0.77 | −1.04;−0.50 |
| West Africa | 0.05 | −0.23;0.33 |
| Age (years) at ART initiation(versus ref‡) | ||
| 18–29 | Ref‡ | |
| 30–34 | −0.60 | −0.83;−0.37 |
| 35–39 | −0.78 | −1.02;−0.54 |
| >39 | −1.16 | −1.38;−0.94 |
| Baseline body weight (per 5 kg)(versus ref‡) | ||
| 55 kg | Ref‡ | |
| For each 5kg higher | −0.34 | −0.38;−0.30 |
| Baseline CD4 count in cells/µl (versus ref‡) | ||
| 0–49 | Ref‡ | |
| 50–199 | 0.20 | −0.03;0.43 |
| >199 | 0.50 | 0.24;0.77 |
Reference group: men aged <30 years who started ART in Southern Africa with a baseline CD4 count <50 cells/µL and with an initial body weight =55 kg.
Patients who started EFV had a higher mean adjusted body weight change in the first year than patients starting NVP (−0.69, 95% CI −0.78; −0.59) or a protease inhibitor (PI) (−1.41, 95% CI −1.79; −1.03) but patients on a PI-based ART regimen had a higher median baseline body weight: 57 kg (IQR 49.6–66) versus 55 kg (IQR 48–62, p=0.0000) compared to patients on a NNRTI regimen. There was no significant difference in CD4 cell count at baseline between patients started on a PI- compared to a NNRTI regimen (117 versus 128 cells/µL, p=0.3502).
In the second year patients initiated on NVP had a greater mean adjusted body weight change than patients started on EFV (0.34, 95% CI 0.23; 0.44) (table 2).
Geographical regions
In the first year mean adjusted body weight change was 1.85 (95% CI 1.54;2.16), 1.56 (95% CI 1.40;1.71) and 0.46 kg/year (95% CI 0.35;0.56) higher in patients from the Central, West and East African regions respectively and 0.41 kg/year (95% CI −0.81; −0.01) lower in the Asia-Pacific region compared to the Southern African region. In the second year, mean adjusted body weight change was lower in patients from the Central African (−0.63, 95% CI −1.19; −0.07), East African (−0.28, 95% CI −0.53; −0.03) and Asia-Pacific (−0.68, 95% CI −1.23; −0.12) regions compared to the Southern African region.
Clinical and biological status
The mean adjusted body weight change in the first year was higher in patients with a poorer clinical status, reflected by higher WHO stage (WHO stage I/II: −2.26 ((95% CI 2.12;2.41) compared to WHO stage IV), lower baseline CD4 lymphocyte count ((baseline CD4 cell count larger than 199: −3.56 (95% CI 3.42;3.70) compared to baseline CD4 cell count 0–49 cells/µL), lower hemoglobin level (baseline hemoglobin ≥10 g/dL: −2.17 (95% CI 1.94;2.41) compared to a hemoglobin <7.5 g/dL) and lower initial body weight (for every 5 kg higher −0.90 (95% CI −0.92; −0.88)). Furthermore mean adjusted body weight change in the first year was higher in patients with a CD4 lymphocyte count change of more than 249 cells/µL (1.43 (95% CI 1.33;1.54)) compared to patients with a CD4 cell count change between 0–149 cells/µL. Contrary to the first year, the mean adjusted body weight change in patients with a poorer clinical state (lower CD4 count, higher WHO stage and lower adjusted body weight) at baseline was lower than in patients with the better clinical state at baseline, but patients with the highest CD4 cell count change (>99 cells/µL) between 12 and 24 months of treatment still had the highest mean adjusted body weight change (0.44, 95% CI 0.33;0.55, compared to 0–49 cells/µL CD4 cell count change).
Weight loss in the second year of treatment
In both the 130,427 patients in the 6–18 month period as in the 84,394 patients in the 18–30 month period with at least 1 weight measurement, the median number of weight measurements was 3 (IQR 2–5). Between 12 and 24 months of ART treatment 45.8% of patients lost weight overall. The median weight loss in this sub-group of patients was 2.0 kg (interquartile range (IQR): 1.0, 3.6 kg). At 24 months of treatment, 68.7% of these patients lost between 0–5% of their body weight at 12 months, 23.4% between 5–10% and 7.9% lost more than 10% of their body weight at 12 months. The median weight gained in the sub-group of patients that gained weight in the second year of therapy was 2.1 kg (IQR: 0.9–4.0). A higher probability for weight loss larger than 5% between 12 and 24 months of treatment (model 2, table 4) was associated with the use of D4T (aOR 1.28, 95% CI 1.19;1.39) or other NRTIs (aOR 1.54, 95% CI 1.22–1.96) compared to TDF and with protease inhibitors (aOR 1.43, 95% CI 1.19;1.72) or NVP (aOR 1.06, 95% CI 1.01–1.12) compared to EFV in initial ART regimen; with women (aOR 1.48, 95% CI 1.41–1.55) compared to men; with a higher baseline bodyweight (aOR 1.05 per 5 kg, 95% CI 1.04;1.06), with a CD4 count change less than 0 in the second year (aOR 1.11, 95% CI 1.04–1.18), with a WHO stage IV and year of ART initiation before 2009 (table 4). A lower probability of weight loss larger than 5% between 12 and 24 months of treatment was associated with West African- (aOR 0.76, 95% CI 0.69–0.82) and East African (0.91, 95% CI 0.86–0.97) regions compared to the Southern African region; age between 35–40 years (aOR 0.88, 95% CI 0.82–0.94) compared to age younger than 29 years and a baseline CD4 cell count above 199 cells (aOR 0.87, 95% CI 0.81–0.94) (table 4).
Table 4.
Adjusted odds ratio (aOR) on weight loss ≥ 5% in the second year of ART of each variable compared to its reference level
| Variables | aOR | 95% CI | p-value | |
|---|---|---|---|---|
| Body weight loss in reference group† | 0.11 | 0.09;0.13 | <0.0001 | |
| Gender (versus ref†) | ||||
| Men | Ref† | |||
| Women | 1.48 | 1.41;1.55 | <0.0001 | |
| NRTI^ in first ART regimen (versus ref†) (p <0.0001) | ||||
| Tenofovir | Ref† | |||
| Stavudine | 1.28 | 1.19;1.39 | <0.0001 | |
| Zidovudin | 0.97 | 0.89;1.05 | 0.420 | |
| Other NRTIs* | 1.54 | 1.22;1.96 | 0.0004 | |
| NNRTI# or PIß in first ART regimen (versus ref†) (p=0.0005) | ||||
| Efavirenz | Ref† | |||
| Nevirapine | 1.06 | 1.01;1.12 | 0.032 | |
| Protease Inhibitor | 1.43 | 1.19;1.72 | 0.0001 | |
| Geographical region (versus ref†)(p=<0.0001) | ||||
| Southern Africa | Ref† | |||
| Asia-Pacific | 0.88 | 0.73;1.06 | 0.169 | |
| Central Africa | 0.82 | 0.67;1.01 | 0.055 | |
| East Africa | 0.91 | 0.86;0.97 | 0.001 | |
| West Africa | 0.76 | 0.69;0.82 | <0.0001 | |
| Year of ART initiation (versus ref†) (p <0.0001) | ||||
| <2005 | 1.23 | 1.03;1.46 | 0.021 | |
| 2005–2006 | 1.38 | 1.18;1.63 | <0.0001 | |
| 2007–2008 | 1.30 | 1.11;1.52 | 0.001 | |
| 2009–2010 | Ref† | |||
| Age (years) at ART initiation(versus ref†) (p= 0.0002) | ||||
| 18–29 | Ref† | |||
| 30–35 | 0.96 | 0.91;1.02 | 0.220 | |
| 35–40 | 0.88 | 0.82;0.94 | 0.0001 | |
| >=40 | 0.99 | 0.94;1.05 | 0.791 | |
| Baseline body weight (per 5 kg)(versus ref†) | ||||
| 55 kg | Ref† | |||
| For each 5kg higher | 1.05 | 1.04;1.06 | <0.0001 | |
| Initial clinical stage (versus ref†) | ||||
| AIDS or WHO IV | Ref† | |||
| CDC B or WHO III | 0.84 | 0.78;0.90 | <0.0001 | |
| CDC A or WHO I/II | 0.93 | 0.87;1.00 | 0.047 | |
| Clinical stage missing | 1.00 | 0.91;1.10 | 0.982 | |
| Baseline CD4 count in cells/µl (versus ref†) | ||||
| 0–9 | Ref† | |||
| 50–199 | 0.96 | 0.90;1.02 | 0.159 | |
| >199 | 0.87 | 0.81;0.94 | 0.0002 | |
| CD4 count change between M12 and M24 in cells/µl (versus ref†) | ||||
| <0 | 1.11 | 1.04;1.18 | 0.002 | |
| 0–49 | Ref† | |||
| 50–99 | 1.03 | 0.97;1.10 | 0.322 | |
| >99 | 0.98 | 0.92;1.04 | 0.546 | |
| Baseline hemoglobin in g/dl (versus ref†) | ||||
| <7.5 | Ref†* | |||
| 7.5–10 | 0.93 | 0.83;1.04 | 0.212 | |
| ≥10 | 0.91 | 0.81;1.02 | 0.095 | |
| Missing | 0.92 | 0.82;1.04 | 0.180 | |
The reference group: baseline body weight=55kg, TDF, EFV, region= Southern Africa, men, WH0 IV, start year 2009–2010, age: 18–30, baseline CD4 cell count <50, cd4change during second year 0–50, hb<7.5
NRTI: Nucleoside/Nucleotide reverse transcriptase inhibitor
Other NRTI’s: didanosine, lamivudine, abacavir and emtricitabine
NNRTI: Non-nucleoside reverse transciptase inhibitor
PI : protease inhibitor
When comparing baseline characteristics between patients with complete data availability and patients where CD4 counts were imputed, results were similar (supplemental tables 3–4).
Discussion
Our study shows that persons with HIV infection experienced the most weight gain during the first year after ART initiation and that the second year weight changes were equally distributed between losses and gains. The study showed that weight gain was more pronounced in patients with a poorer clinical condition at the start of ART, reflected by a low pre-treatment CD4 cell count, a more advanced WHO clinical stage, lower hemoglobin level and lower baseline body weight. This has also been reported previously.9, 10 Hurley et al also reported that the lower the BMI of patients was at start of ART, the higher the weight gain was after one year of treatment in a single cohort in South Africa.11
Almost half of the patients lost weight in the second year of treatment, with almost 8% of these patients losing more than 10% of their weight at 12 months of treatment. It might be due, among other reasons, to therapy failure, opportunistic infection (tuberculosis) or intentional weight loss (diet or exercise). A study from Rwanda also showed that with D4T based ART a high proportion of patients had a progressive decline of body weight in the second year of treatment.12 There is not enough data available in our study to fully explain this weight loss. There was however a difference in total weight change after 2 years of treatment according to the antiretroviral drugs in the initial regimen, with in the NRTI-category the highest weight gain in patients on a TDF-containing regimen. Certain patients on a D4T regimen had a smaller weight gain or even lost weight after one year of ART. This has also been observed in a study from Rwanda where such weight loss was probably due to lipoatrophy, a consequence of mitochondrial toxicity.13 In our analysis adjusted for CD4 cell count change over the 12–24 month period, the smaller weight gain in patients on a D4T-containing regimen was still seen. Indeed, not only treatment failure but also adverse events such as painful neuropathies can contribute to weight loss. Others also showed that the impact of virological failure on weight evolution during the second year of ART was not significant.11 Although lipoatrophy measurements were not available in our database, our hypothesis is that this D4T related weight loss is indeed due to lipoatrophy. The lower weight gain in patients on D4T was mainly seen in patients from East-Africa, females, patients with a higher baseline body weight, older patients and patients with a CD4 cell count below 50 cells/µL at baseline. This corresponds to the findings of van Griensven et al, who showed that being female, a longer duration of ART, a baseline BMI ≥25 kg/m2 a D4T-based ART-regimen (compared to an AZT-based regimen), but also (in contrast to our findings) a baseline CD4 cell count ≥150 cells/µL, were significantly associated with lipoatrophy.12 A South African study showed that lactic acidosis, another consequence of mitochondrial toxicity, was also associated with female gender, increased BMI and D4T therapy.14
Our study showed, in contrast to other studies15 that patients started on a protease inhibitor (PI)-based therapy (mostly boosted lopinavir) had on average a lower body weight increase than patients starting on a NNRTI-based regimen. The group of patients started on a PI-based ART regimen seems to be a diverse group; in West Africa patients on PI-based therapy have a lower CD4 cell count than those not on PIs, while in Asia-Pacific the opposite is the case. Patients who started a PI-boosted ART regimen might have been patients previously exposed to a NNRTI, e.g. because of pregnancy (and therefore already had a higher CD4 cell count), or were patients from more wealthy Asian countries where PIs may be used as first line therapy and at higher CD4 count. Excluding from the analysis patients with HIV2 infection on PIs, did not change the results.
Although weight gain is mostly regarded as being beneficial in patients with HIV, we have to take into account the current obesity epidemics in several resource limited settings.16 Overweight and obesity are indeed important health risk factors predisposing for cardiovascular and metabolic diseases and certain cancer types.17 Overweight and obese patients with HIV were found to have a higher prevalence of multi-morbidity than patients with normal or underweight. 18
The strength of our study lies in its large number of participants from various resource-limited countries. This is a strong advantage to detect statistical significant differences between treatment regimens. However, it is unclear how clinically significant some of these differences are. Indeed, our study had several limitations. HIV treatment centers participating in the IeDEA network cannot be considered to be representative of all the HIV treatment centers in the regions included in the study. The majority of patients were included in the Southern African region (67.7%) (with almost 75% of patients coming from Zambia) and only 0.80% from the Asia-Pacific region. We did however correct the outcomes for region, but cannot prevent some bias of data. Moreover, data were collected during routine care and therefore were not of the same quality as data collected during clinical trials. Our cohort data contained many missing variables and no information was available on adherence to drugs, drug switches or discontinuations, co-infections, lipodystrophy, access to food programs, pregnancies during follow up and number of patients who were lost to follow up, transferred or who died. HIV RNA plasma viral load measurements were available in a subgroup of patients. However, in some regions of all measured viral loads the majority (up to 96.1% in the East African region) was 50 copies/mL or higher at month 24 after initiation of ART. This suggests that viral load measurements were only performed if there were clinical clues of failure (or non-adherence).
In conclusion, body weight increase, particularly in the first year of ART, is correlated with CD4 count increase and therefore may suggest that a patient is on an effective ART regimen. After the first year of ART the body weight increase is leveling off and in certain patients on a D4T regimen body weight may even decrease because of lipoatrophy. Hopefully with the 2013 WHO treatment guidelines, where D4T is not advised as the preferred first line therapy, and with patients starting ART much earlier, there will be less thymidine-related side effects.14 Weight and ideally BMI measurements may then become more meaningful to monitor treatment response from a clinical standpoint as well as basic program success indicator.
Supplementary Material
Acknowledgments
We are grateful to all patients, care givers and data managers involved in the participating cohorts and treatment programmes. This work was supported by a contract provided by the Joint United Nations Programme on HIV/AIDS (UNAIDS).
Funding:
DH: FWO Research Foundation Flanders, Brussels, Belgium
EB/ME/AJ: The National Cancer Institute (NCI), the Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD) and the National Institute of Allergy and Infectious Diseases (NIAID) of the U.S. National Institutes of Health (NIH), as part of the International Epidemiologic Databases to Evaluate AIDS (IeDEA) support to West Africa Network under Award Number U01AI069919. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
LD: National Institute of Health (NIH): grant number UO1AI069911-08
MF: National Institute of Allergy and Infectious Diseases [K01AI083097] and by the generous support of the American people through Cooperative Agreement AID 674-A-12-00029 from the United States Agency for International Development (USAID). The contents are the responsibility of the authors and do not necessarily reflect the views of USAID or the United States Government.
CH: National Institute of Allergy and Infectious Diseases, 083099
Footnotes
Competing interests: the authors declare that they have no conflicts of interest.
References
- 1.Loveday M, Scott V, McLoughlin J, et al. Assessing care for patients with TB/HIV/STI infections in a rural districs in KwaZulu-Natal. S Afr Med J. 2011;101(12):887–890. [PubMed] [Google Scholar]
- 2.Colebunders R, Moses KR, Laurence J, et al. A new model to monitor the virological efficacy of antiretroviral treatment in resource-poor countries. Lancet Infect Dis. 2006 Jan;6(1):53–59. doi: 10.1016/S1473-3099(05)70327-3. [JIF:13.165] [DOI] [PubMed] [Google Scholar]
- 3.Range NS, Malenganisho W, Temu MM, et al. Body composition of HIV-positive patients with pulmonary tuberculosis: a cross-sectional study in Mwanza, Tanzania. Ann Trop Med Parasitol. 2010;104(1):81–90. doi: 10.1179/136485910X12607012373830. [DOI] [PubMed] [Google Scholar]
- 4.Madec Y, Szumilin E, Genevier C, et al. Weight gain at 3 months of antiretroviral therapy is strongly associated with survival: evidence from two developing countries. AIDS. 2009;23(7):853–861. doi: 10.1097/QAD.0b013e32832913ee. [DOI] [PubMed] [Google Scholar]
- 5.Boulle A, Orrel C, Kaplan R, et al. International Epidemiological Databases to Evaluate Aids in Southern Africa Collaboration. Substitutions due to antiretroviral toxicity or contraindication in the first 3 years of antiretroviral therapy in a large South African cohort. Antivir Ther. 2007;12(5):753–760. doi: 10.1177/135965350701200508. [DOI] [PubMed] [Google Scholar]
- 6.Koethe JR, Lukusa A, Giganti MJ, et al. Association between weight gain and clinical outcomes among malnourished adults initiating antiretroviral therapy in Lusaka, Zambia. J Acquir Immune Defic Syndr. 2010;53(4):507–513. doi: 10.1097/QAI.0b013e3181b32baf. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sudfeld CR, Isanaka S, Mugusi FM, et al. Weight change at 1 mo of antiretroviral therapy and its association with subsequent mortality, morbidity and CD4 T cell reconstitution in a Tanzanian HIV-infected adult cohort. Am J Clin Nutr. 2013;97(6):1278–1287. doi: 10.3945/ajcn.112.053728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Egger M, Ekouevi DK, Williams C, et al. Cohort Profile: The international epidemiological databases to evaluate AIDS (IeDEA) in sub-Saharan Africa. Int J Epidemiol. 2012;41:1256–1264. doi: 10.1093/ije/dyr080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lakey W, Yang LY, Yancy W, et al. Short communication: from wasting to obesity: initial antiretroviral therapy and weight gain in HIV-infected persons. AIDS Res Hum Retroviruses. 2013 Mar;29(3):435–440. doi: 10.1089/aid.2012.0234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tate T, Willig AL, Willig JH, et al. HIV infection and obesity: where did all the wasting go? Antivir Ther. 2012;17(7):1281–1289. doi: 10.3851/IMP2348. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hurley E, Coutsoudis A, Giddy J, et al. Weight evolution and perceptions of adults living with HIV following initiation of antiretroviral therapy in a South African urban setting. S Afr Med J. 2011;101(9):645–650. [PubMed] [Google Scholar]
- 12.van Griensven J, Zachariah R, Mugabo J, et al. Weight loss after the first year of stavudine-containing antiretroviral therapy and its association with lipoatrophy, virological failure, adherence and CD4 counts at primary health care level in Kigali, Rwanda. Trans R Soc Trop Med Hyg. 2010;104(12):751–757. doi: 10.1016/j.trstmh.2010.08.016. [DOI] [PubMed] [Google Scholar]
- 13.van Griensven J, De Naeyer L, Mushi T, et al. High prevalence of lipoatrophy among patients on stavudine-containing first-line antiretroviral therapy regimens in Rwanda. Trans R Soc Trop Med Hyg. 2007 Aug;101(8):793–798. doi: 10.1016/j.trstmh.2007.02.020. [DOI] [PubMed] [Google Scholar]
- 14.Dlamini J, Ledwaba L, Mokwena N, et al. Lactic acidosis and symptomatic hyperlactataemia in a randomized trial of first-line therapy in HIV-infected adults in South Africa. Antivir Ther. 2011;16(4):605–609. doi: 10.3851/IMP1790. [DOI] [PubMed] [Google Scholar]
- 15.Nguyen A, Calmy A, Schiffer V, et al. Lipodystrophy and weight changes: data from the Swiss HIV Cohort Study, 2000–2006. HIV Med. 2008;9(3):142–150. doi: 10.1111/j.1468-1293.2007.00537.x. [DOI] [PubMed] [Google Scholar]
- 16.World Health Organisation. [accessed 01/02/2015];Overweight/obesity by country. Available on: http://apps.who.int/gho/data/node.main.A897?lang=en.
- 17.World Health Organization. Global health risks. Mortality and burden of disease attributable to selected major risks. 2009 ISBN 978 92 4 156387.
- 18.Kim DJ, Westfall AO, Chamot E, et al. Multimorbidity patterns in HIV-infected patients: the role of obesity in chronic disease clustering. J Acquir Immune Defic Syndr. 2012;16(5):600–605. doi: 10.1097/QAI.0b013e31827303d5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.